Assessing visible aerosol generation during vitrectomy in the era of Covid-19

Objective To assess visible aerosol generation during simulated vitrectomy surgery. Methods A model comprising a human cadaveric corneoscleral rim mounted on an artificial anterior chamber was used. Three-port 25 gauge vitrectomy simulated surgery was performed with any visible aerosol production recorded using high-speed 4K camera. The following were assessed: (1) vitrector at maximum cut rate in static and dynamic conditions inside the model, (2) vitrector at air-fluid interface in a physical model, (3) passive fluid-air exchange with a backflush hand piece, (4) valved cannulas under air, and (5) a defective valved cannula under air. Results No visible aerosol or droplets were identified when the vitrector was used within the model. In the physical model, no visible aerosol or droplets were seen when the vitrector was engaged at the air-fluid interface. Droplets were produced from the opening of backflush hand piece during passive fluid-air exchange. No visible aerosol was produced from the intact valved cannulas under air pressure, but droplets were seen at the beginning of fluid-air exchange when the valved cannula was defective. Conclusions We found no evidence of visible aerosol generation during simulated vitrectomy surgery with competent valved cannulas. In the physical model, no visible aerosol was generated by the high-speed vitrector despite cutting at the air-fluid interface.